Factors influencing the chemistry of the near-field Columbia River plume: Nitrate, silicic acid, dissolved Fe, and dissolved Mn

Transcription

1 Click Here for Full Article JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 113,, doi: /2007jc004702, 2008 Factors influencing the chemistry of the near-field Columbia River plume: Nitrate, silicic acid, dissolved Fe, and dissolved Mn Kenneth W. Bruland, 1,2 Maeve C. Lohan, 3 Ana M. Aguilar-Islas, 4 Geoffrey J. Smith, 1 Bettina Sohst, 1 and Antonio Baptista 5 Received 2 January 2008; revised 19 September 2008; accepted 2 October 2008; published 30 December [1] Factors influencing concentrations of nitrate, silicic acid, dissolved Fe, and dissolved Mn in the near-field Columbia River plume were examined during late spring and summer from 2004 to 2006 as part of the River Influences on Shelf Ecosystems program. Under upwelling-active phases, cold, high-nitrate coastal seawater was entrained in the plume, and nitrate concentrations of mm were observed with as much as 90% from a coastal seawater origin. Under downwelling-relaxation phases, warm, nutrient-depleted coastal seawater was entrained forming a near-field plume with nitrate concentrations of mm, with the river as the only source. Elevated silicic acid in the river is the dominant source, with concentrations of mm in the near-field plume. During upwelling-active phases, high concentrations of dissolved Fe (as high as 40 nm) in the cold, low-oxygen, nutrient-rich coastal seawater were entrained to form a near-field plume with nm dissolved Fe. During downwelling-relaxation phases, dissolved Fe in the intruding underlying warm coastal seawater was 1 3 nm, producing plume concentrations of 2 13 nm, with higher concentrations during the high river flow of May Dissolved Mn in the near-field plume covaried markedly as a function of increased tidal flushing in the estuary. The use of CORIE (pilot environmental observation and forecasting system for the Columbia River) time series conductivity-temperature-depth data within the estuary, along with data presented in this study, allows extrapolation of the near-field plume chemistry throughout the spring and summer seasons to provide insight into this important source of nutrients to the coastal waters in this region. Citation: Bruland, K. W., M. C. Lohan, A. M. Aguilar-Islas, G. J. Smith, B. Sohst, and A. Baptista (2008), Factors influencing the chemistry of the near-field Columbia River plume: Nitrate, silicic acid, dissolved Fe, and dissolved Mn, J. Geophys. Res., 113,, doi: /2007jc Introduction [2] The Columbia River is the largest river entering the eastern boundary of the North Pacific Ocean [Thomas and Weatherbee, 2006], and during the summer it contributes 90% of the freshwater entering the California Current system along the U.S. west coast between the Strait of Juan de Fuca and San Francisco Bay [Barnes et al., 1972]. The Columbia River plume is an important feature off the Washington and Oregon coasts, and is characterized by a shallow (2 20 m) surface lens of low-salinity water. During sustained upwelling conditions this buoyant plume 1 Institute of Marine Sciences, University of California, Santa Cruz, California, USA. 2 Department of Ocean Sciences, University of California, Santa Cruz, California, USA. 3 School of Earth, Ocean, and Environmental Sciences, University of Plymouth, Plymouth, UK. 4 International Arctic Research Center, University of Alaska, Fairbanks, Alaska, USA. 5 Department of Science and Engineering, Oregon Health and Science University, Beaverton, Oregon, USA. Copyright 2008 by the American Geophysical Union /08/2007JC004702$09.00 tends to move offshore and southward, while during downwelling conditions the plume moves northward and nearshore forming a narrow coastal jet [Landry et al., 1989; Hickey et al., 2005]. During wind reversals, the plume can be bidirectional [Hickey et al., 2005, 2008]. With a shift from equatorward upwelling winds to poleward downwelling winds, the southwest plume moves onshore over the Oregon shelf concurrent with the rapid formation of a nearshore northward flowing plume. In contrast, with the onset of upwelling favorable winds, the northward plume advects and mixes offshore over the Washington shelf while a southwest flowing plume is rapidly initiated. [3] The Columbia River plume is an important source of macro- and micronutrients to the coastal waters off Washington and Oregon [Hill and Wheeler, 2002; Lohan and Bruland, 2006; Aguilar-Islas and Bruland, 2006], thus directly impacting the lowest trophic levels (R. M. Kudela and T. D. Peterson, Influence of a buoyant river plume on phytoplankton nutrient dynamics: What controls standing stocks and productivity?, submitted to Journal of Geophysical Research, 2008). The near-field Columbia River plume entering the coastal waters generally has a salinity of 10 25, composed of roughly two-thirds to one-fourth Columbia 1of23

2 Figure 1. Location of stations used in this study. Surface transect across the near-field plume is indicated by the gray circle. RISE estuarine stations are indicated by the gray diamonds. CORIE CTD stations are indicated by the black diamonds. RISE river stations are indicated by the white circles. USGS NASQAN river station is indicated by the white square. River water and one-third to three-fourths coastal seawater, varying with the magnitude of both tidal mixing and freshwater river input (J. D. Nash et al., Turbulent mixing in the Columbia River Estuary: Structure and consequences for plume composition, submitted to Journal of Geophysical Research, 2008). The entrainment and mixing of this coastal seawater with Columbia River water to form the near-field plume takes place both within, and at or near the mouth of the estuary [Barnes et al., 1972; Nash et al., submitted manuscript, 2008]. The chemistry of the river water coastal seawater, and estuarine water that mix to form the plume vary temporally as a function of season, oceanographic conditions, tidal phase and river flow. As a result, the chemistry of the near-field plume can vary markedly with changes in these conditions Columbia River Freshwater End-Member [4] The character of the Columbia River varies on a seasonal and interannual basis. Daily values of river discharge and approximately monthly sampling of temperature, nitrate and silicic acid at the Beaver Army Terminal station (RM53), near Quincy, Oregon (Figure 1) over the last decade are reported by the U.S. Geologic Survey s National Stream Water Quality Network ( gov/nasqan/data/finaldata/beaver.html) and are presented in Figure 2. The temperature of the Columbia River increases through the spring to a maximum of C in July and August (Figure 2b). Nitrate concentrations in the Columbia River (Figure 2a) are at a maximum in winter (concentrations up to 50 mm), coincident with high winter rainfall and high flow from coastal tributaries draining the coastal subbasin joining the Columbia River west of the Cascade mountain range. Nitrate concentrations decrease markedly through April and May, and usually reach a minimum of 2 10 mm in June and July (Figure 2a). The silicic acid concentration in the river is high all year ( mm), with minimum concentrations ( mm) in the summer months (Figure 2b). The Columbia River is relatively unique for major rivers in being silicic acid rich and relatively nitrate poor during the summer months, with silicic acid:nitrate ratios ranging from 10 to 50. In contrast, the Mississippi River and other major rivers in North America and Europe have become nitrate rich because of anthropogenic inputs of fixed nitrogen, and their silicic acid:nitrate ratio has dropped to 1 because of the marked increase in nitrate [Cloern, 2001] Coastal Seawater End-Member [5] The coastal seawater at the mouth of the Columbia River is mixed with river water within and/or near the mouth of the estuary to form the near-field Columbia River plume [Barnes et al., 1972; Jay et al., 1990; Nash et al., submitted manuscript, 2008; E. D. Zaron and D. A. Jay, Mixing in the tidal plume of the Columbia River, submitted to Journal of Geophysical Research, 2008]. The seawater source mixing to form the near-field plume is found at depths of 5 20 m near the mouth of the estuary in the core of the entering salt wedge or underlying seawater. Fluctuations in the properties of this coastal seawater end-member correspond to fluctuations in the strength and persistence of regional-scale synoptic winds and local coastal upwelling activity [Hickey et al., 2006]. The upwelling-favorable, equatorward, wind stress can impact the chemistry of the plume by bringing cold, higher-salinity, low-oxygen, nutrient-rich water up the shelf to shallow depths at the mouth of the estuary to be entrained with the river water. Stefansson and Richards [1963] and Hickey et al. [2008] have reported that such upwelling conditions are an important process adding nutrients to the surface waters off the northern Washington coast because of entrainment of nutrient-rich subsurface waters by the estuarine-like circulation in the Straits of Juan de Fuca. Monteiro and Largier [1999], in a study of the Saldanha Bay in the Benguela upwelling system, have referred to this upwelling situation supplying cold, nutrient-rich water to be entrained into the Bay as an active phase. The reverse situation was termed the relaxation phase. In this situation, downwelling favorable winds can result in warm, lower-salinity, nutrient-depleted coastal surface waters being observed at depths of 5 20 m at the mouth of the estuary. [6] This variation of the coastal seawater end-member, alternating back and forth between cold, nutrient-rich, higher-salinity waters during upwelling-active phases and warm, nutrient poor, lower-salinity waters during downwelling-relaxation phases, can be ecologically critical in supplying macro and micronutrients to the Columbia River plume (Kudela and Peterson, submitted manuscript, 2008). In addition to influencing the supply of nitrate and silicic acid to the plume [Lohan and Bruland, 2006; Aguilar-Islas and Bruland, 2006], variations in the coastal seawater source can influence dissolved Fe concentrations [Lohan and Bruland, 2006]. [7] The focus of this study is to examine how these changes in the chemistry of river and coastal water endmembers, together with the tidal conditions and river 2of23

3 Figure 2. Data from the US Geological Survey s National Stream Water Quality Network at the Beaver Army Terminal station near Quincy, Oregon. Daily data on (a) river discharge (in gray) and approximately monthly data on nitrate (in black) and (b) temperature (in gray) and silicic acid (in black). discharge, can influence the macronutrient, and dissolved Fe and Mn chemistry of the near-field Columbia River plume. The coastal ocean processes River Influence on Shelf Ecosystems (RISE) program provided a unique opportunity to examine the formation of the near-field plume during varying conditions in late spring and summer periods of In addition to the selective observations made during the RISE cruises, the continuous time series of temperature and salinity from conductivity-temperaturedepths (CTDs) located at 7 m depth at CORIE (pilot environmental observation and forecasting system for the Columbia River) stations provides information from within the estuary that can be used as an indication of the local upwelling conditions and to predict likely plume nutrient characteristics during longer temporal scales. CORIE [Baptista, 2006] is an end-to-end observatory for the Columbia River estuary and plume, with an extensive observation network of physical parameters coupled with a semioperational modeling system [Baptista et al., 2005; Y. J. Zhang et al., Daily forecasts of Columbia River plume circulation: A tale of spring/summer cruises, submitted to Journal of Geophysical Research, 2008], both served by and integrated through flexible cyberinfrastructure [Howe et al., 2007; L. Bright, D. Maier, and B. Howe, Managing the forecast factory, Proceedings of the 22nd ICDE Workshop on Workflow and Data Flow for Scientific Applications, 2006, available at factory_final.pdf]. 2. Methods 2.1. Sample Collection [8] Sampling off the coasts of Washington and Oregon was accomplished during five research cruises as part of the collaborative RISE program. The pre-rise cruise was aboard the R/V Point Sur and took place from 27 June to 3of23

4 Figure 3a. Wind vector plots at the NDBC Buoy located just off the mouth of the Columbia River in the vicinity of the near-field plume surface transect. The stick plots depict both the wind speed and direction (positive direction is poleward/downwelling and negative direction is equatorward/ upwelling) for the months of May August for July All other research cruises were on board the R/V Wecoma, with RISE-1W from 8 to 23 July 2004, RISE-2W from 29 May to 20 June 2005, RISE-3W from 4 to 26 August 2005, and RISE-4W from 21 May to 11 June Surface samples were collected with a clean surface pump fish system [Bruland et al., 2005] equipped with a YSI 600 OMS CTD sonde. The sonde was calibrated against the ship s calibrated Sea-Bird Electronics SBE 911plus Conductivity, Temperature, and Depth system [Hickey et al., 2008]. Samples for dissolved trace metals were filtered in-line through acid-cleaned 0.45 mm pore size Teflon 2 membrane polypropylene capsule filters (GE Osmonics) [Bruland et al., 2005]. Vertical profiles down to approximately 20 m were collected with the fish, and deeper samples were collected with Teflon 2 coated GO-Flo samplers (General Oceanics) deployed on Kevlar 2 hydroline [Bruland et al., 1979] Analytical Methods [9] Macronutrients (nitrate + nitrite (referred to herein as nitrate), silicic acid and phosphate) were measured on a Lachat QuikChem 8000 Flow Injection Analysis system using standard colorimetric methods [Parsons et al., 1984]. Samples for the determination of dissolved Fe and Mn were acidified to ph using subboiled quartz distilled 6 N hydrochloric acid (Q-HCl) (using the equivalent of 4 ml acid per liter of seawater; M HCl), and were allowed to sit for at least 30 min after acidification prior to analysis. Dissolved Fe and Mn were determined onboard ship by flow injection (FI) methods involving in-line preconcentration, catalytic enhancement with Fe or Mn acting as a catalyst for the formation of a colored end product, and spectrophotometric detection. Details are found elsewhere (Fe [Lohan et al., 2006] and Mn [Aguilar-Islas et al., 2006]). [10] Analytical figures of merit for the various techniques include estimates of accuracy within 4% for dissolved Fe 4of23

5 Figure 3b. Wind vector plots at the NDBC Buoy located just off the mouth of the Columbia River in the vicinity of the near-field plume surface transect. The stick plots depict both the wind speed and direction (positive direction is poleward/downwelling and negative direction is equatorward/ upwelling) for the months of May August for based upon analyses of SAFe standards, within 3% for dissolved Mn based upon analysis of NASS-4, and within 4%, 3%, and 5% for nitrate (+nitrite), phosphate and silicic acid (respectively) on MOOS-1, CRM (NRC). Estimates of precision are on the order of 5% for dissolved Fe and Mn, and 2% for macronutrients. Detection limits are 0.02 nm dissolved Fe, 0.5 nm dissolved Mn, 0.05 mm nitrate (+nitrite), 0.03 mm phosphate and 0.2 mm silicic acid. 3. Results [11] Results from the five RISE cruises are presented in a chronological fashion starting with the pre-rise cruise in early July 2004 and ending with the RISE-4W cruise in May The results presented herein consist of data collected in the river, the estuary, and coastal waters in the vicinity of the near-field plume, both within the near-field plume and just outside it (see Figure 1 for station locations). The time series data for winds (Figures 3a 3c) and temperature and salinity from stations within the estuary (e.g., Figures 4, 7, and 11) provide a longer-term perspective (mid-may August of each year) to the cruise data collected during selective short time periods of these months. Wind directions and velocities at the Columbia River buoy (B46029) located just offshore of the mouth of the Columbia River for the months of May August for provide a measure of the local wind stress and are generally consistent with the regional forcing in this area (Figures 3a 3c). The stick plots in the negative direction are equatorward, upwelling favorable winds, while the winds in the positive direction are poleward, downwelling favorable. The temperature of the river increases from May to July, and the maximum temperature observed in the daily estuary time series reflects this increased temperature of the river end-member. There is also a neap/spring tidal cycle with slightly cooler daily temperatures (and higher salinities) observed at these lower estuary stations during spring tides Data From 2004 [12] A time series of temperature data from the estuarine station at Jetty A for mid-may through the end of August 2004 (Figure 4) indicated that an upwelling-active 5of23

6 Figure 3c. Wind vector plots at the NDBC Buoy located just off the mouth of the Columbia River in the vicinity of the near-field plume surface transect. The stick plots depict both the wind speed and direction (positive direction is poleward/downwelling and negative direction is equatorward/ upwelling) for the months of May August for phase existed during July, with the exception of a short downwelling-relaxation phase during July. Temperature/salinity plots corresponding to our two sampling periods (2 and 21 July) are also presented in Figure 4. A period of strong upwelling favorable winds from the north during 28 June to 2 July preceded the 2 July pre-rise sampling date and Figure 4 shows that the coastal seawater mixing with the Columbia River water was cold (9 C) with a high salinity (33.5). [13] A period of strong downwelling favorable winds took place on July 16 and again from 18 to 21 July (Figure 3a). The temperature of the coastal seawater mixing to form the plume rose markedly starting on 18 July (e.g., the rise in daily low temperature within the estuary as shown in Figure 4) with a corresponding drop in salinity, and by 20 July a downwelling-relaxation phase had occurred where the temperature of the coastal seawater being mixed with the river water increased to14 C and the salinity decreased to On 21 July the winds had switched back to be from the north and by 22 July an upwelling-active phase once again prevailed with cold, higher-salinity coastal seawater mixing with the river water to form the plume. Cruise data from the estuary and near-field plume were obtained on 2 July during the pre-rise cruise and on 21 July during RISE-1W and correspond to the dates with temperature/salinity plots shown in Figure 4. [14] The pre-rise cruise data from 2 July 2004 are representative of the upwelling-active phase prevalent during most of the month of July 2004 and were collected during an intense spring tide. Figure 5a shows that river water of 20.5 C was mixing with cold (8.5 to 9.0 C) coastal seawater intruding into the estuary to form a plume of salinity 20 and a temperature of C. In this case 61% of the plume was coastal seawater, and 39% was Columbia River water. Figure 5b shows that silicic acid-rich Columbia River water with a concentration of 155 mm was mixing with cold, nutrient-rich seawater with 35 mm silicic acid to form a plume of salinity 20 with 80 mm silicic acid. The plume had 74% of its silicic acid from the river source and 26% from the seawater source. The plot of salinity versus nitrate 6of23

7 (Figure 5c) shows river water with only a few mm nitrate mixing with cold, nutrient-rich upwelled water with 26 mm nitrate to form a plume with a salinity of 20 and a nitrate concentration of 16 mm. In this case, 95% of the elevated nitrate in the plume had a coastal seawater source associated with the upwelling-active conditions and only 5% came from the Columbia River. [15] The upwelling-active phase during and preceding the pre-rise 2 July sampling period, together with the spring tide conditions, resulted in the on-shelf benthic boundary layer transport of low temperature, higher-salinity, lowoxygen, high-nutrient, and high-iron water up onto the inner shelf to depths of 5 20 m near the mouth of the estuary, and it was this upwelled water that was mixing with the Columbia River to form the plume. Figure 5d shows the elevated dissolved iron concentrations of nm in this near bottom water being mixed with relatively low dissolved iron water within the estuary. Dissolved Fe behaves nonconservatively within the low-salinity region of the estuary and there is evidence of flocculation and removal of dissolved Fe coming from the river source [Buck et al., 2007]. The plume water at a salinity of 20 has 16 nm dissolved iron that appears to be coming primarily from the coastal seawater source in this upwelling-active phase. The dissolved Mn concentrations in the plume were 200 nm, values much greater than observed in either the river or lower-salinity regions of the estuary, or in the cold, nutrientrich seawater forming the salt wedge (Figure 5e). Dissolved Mn also behaves nonconservatively within the estuary, but in contrast to dissolved Fe, there is evidence of an estuarine input of dissolved Mn that correlates with tidal amplitude [Aguilar-Islas and Bruland, 2006]. It appears that under intense spring tide conditions the added turbulence within the estuary results in elevated dissolved Mn. [16] The RISE-1W data from 21 July 2004 are from a downwelling-relaxation phase during a neap tide cycle. Figure 6a illustrates that river water of 22 C was mixing with warm (15 C) coastal seawater to form a plume of salinity 20 with a temperature of C. At a salinity of 20, 60% of the plume was coastal seawater and 40% was Columbia River water. Figure 6b shows that river water of 160 mm silicic acid was mixing with warm, nutrient depleted seawater with <3 mm silicic acid to form a plume of salinity 20 with 60 mm silicic acid. In this case the plume obtained essentially all of its silicic acid from the river source. The plot of salinity versus nitrate (Figure 6c) shows river water with 5 mm nitrate mixing with warm, nutrient depleted water with <1 mm nitrate to form a plume with a salinity of 20 and a nitrate concentration of 3 mm. In this case the source of the low concentration of nitrate present in the plume was from the Columbia River. [17] The downwelling-relaxation phase during the RISE- 1W 21 July sampling period meant that warm, lowersalinity, nutrient-depleted coastal seawater with relatively low dissolved Fe concentrations was being drawn into the Columbia River estuary to mix and form the plume. Figure 6d shows low dissolved iron concentrations of only 1 or 2 nm in this coastal water at depths of 5 20 m being mixed with low dissolved iron water within the estuary. Again, dissolved Fe behaves nonconservatively within the low-salinity region of the estuary, and there is evidence of removal of dissolved Fe coming from the river source. The plume water at a salinity of 20 has only 2 or 3 nm dissolved Fe. These concentrations are an order of magnitude less than the dissolved iron concentrations observed in the plume during the upwelling-active phase of the pre-rise cruise. The dissolved Mn concentrations in the plume were 50 nm, a concentration much lower than observed earlier in the month, but much greater than observed in the warm, nutrient-poor coastal seawater and similar to the concentrations within the low-salinity regions of the estuary (Figure 6e). It appears that, under these lower-energy neap tide conditions, the estuary was a much reduced source of dissolved Mn to the plume Data From 2005 [18] Cruise data were obtained on 5 7 June and June during the RISE-2W cruise, and on August during RISE-3W. The period in 2005 from mid-may to mid-july was characterized as a downwelling-relaxation phase with warm surface water mixing with the river water (Figure 7). During these two months there were sustained periods of either winds from the south or low winds resulting in the continuous downwelling-relaxation phase (Figures 3b and 7). On 14 July 2005 strong winds from the north picked up and were sustained almost continuously through the rest of July and into mid-august. Within a few days after the initiation of these strong winds from the north, cold, higher-salinity water was observed to be mixing with the river water and the system had entered into an upwelling-active phase. [19] The RISE-2W data from 5 to 7 June 2005 are within the long-term downwelling-relaxation phase and during an intermediate tidal cycle. During this sampling interval the winds were from the south (Figure 3b). Although there had been a few days preceding this sampling period with relatively weak winds from the northwest, it was not enough for the system to move all the way into an upwelling-active phase (Figure 7). Figure 8a shows that river water of 16 C was mixing with warm (14 C), low-salinity (30) coastal seawater to form a plume of salinity and a temperature of 15 C. Figure 8b shows that river water of 205 mm silicic acid was mixing with warm, nutrient depleted seawater with <3 mm silicic acid to form a plume that at a salinity of 15 had a concentration of 103 mm silicic acid. The plot of salinity versus nitrate (Figure 8c) shows river water with mm nitrate mixing with warm, nutrient depleted water with <1 mm nitrate to form a plume with a salinity of 15 and a nitrate concentration of 9 mm. In this case, the nitrate concentration in the river was anomalously high compared with the historical data (Figure 2a) and most of the nitrate in the plume had a river source. [20] Relatively low dissolved Fe concentrations of 3 5 nm in subsurface waters were being mixed with relatively low dissolved Fe concentrations within the estuary (Figure 8d). The plume water at a salinity of 15 had 8 12 nm dissolved iron. The dissolved Mn concentrations in the plume ranged from 50 to 100 nm with a value of 80 nm at a salinity of 15. These dissolved Mn concentrations in the plume were greater than those observed in the warm, nutrient poor seawater and also higher than in the lower-salinity regions of the estuary (Figure 8e). [21] The RISE-2W data from 12 to 13 June 2005 are also from a downwelling-relaxation phase, but coincided with a 7of23

8 Figure 4. A time series of temperature for a CTD located at 7 m depth just inside the Columbia River estuary at the Jetty A CORIE station for mid-may August In addition, temperature-salinity diagrams of the CORIE data for 2 relevant days (2 and 21 July) corresponding to the field studies are also presented. The triangles represent a coastal station within 20 km of the river mouth at close to the same time. neap tidal cycle. Figure 9a shows that river water of 16 C was mixing with warm (14 C), low-salinity (30) seawater to form a plume with a salinity of and a temperature of 15 C. Figure 9b shows that river water of 190 mm silicic acid was mixing with warm, nutrient depleted seawater with <5 mm silicic acid to form a plume with 100 mm silicic acid at a salinity of 15. The plot of salinity versus nitrate (Figure 9c) shows river water with 13 mm nitrate mixing with warm, nutrient depleted water with <1 mm nitrate to form a plume with a salinity of 15 and a nitrate concentration of 7 mm, with nitrate in the plume coming almost exclusively from the river source. Again, this data shows that the elevated nitrate concentrations in the Columbia 8 of 23

9 Figure 5. A series of property salinity plots for data collected on the pre-rise cruise on 2 July 2004 within the estuary (gray diamonds), within the core of the near-field plume (gray circles), and in a vertical profile of nearby coastal seawater (dissolved Fe and Mn data are for the subsurface coastal seawater end-member only (white triangles)). This was during an upwelling-active phase and cold, relatively high-salinity, nutrient-rich coastal seawater was observed at depths of approximately 10 m and was the water entrained together with the river water to form the plume. 9of23

10 Figure 6. A series of property salinity plots for data collected on the RISE-1W cruise on 21 July 2004 within the estuary (gray diamonds), within the core of the near-field plume (gray circles), and in a vertical profile of nearby coastal seawater (white triangles). This was during a downwelling-relaxation phase and warm, relatively low-salinity, nutrient-depleted coastal seawater was observed at depths of approximately 10 m and was the water entrained together with the river water to form the plume. River in the summer of 2005 were anomalously high compared to previous years (Figure 2). [22] Figures 9d and 9f (with increased vertical axis to display all data) show low dissolved iron concentrations of only a few nm in the coastal water at depths of 5 20 m being mixed with relatively low dissolved iron water within the estuary. The plume water at a salinity of 15 had 6 7 nm dissolved iron. The dissolved Mn concentrations in the plume ranged from 15 to 20 nm with a value of 16 nm at a salinity of 15. These low dissolved Mn concentrations observed in the plume were lower than observed in the warm, nutrient-poor coastal seawater, and approximately 10 of 23

11 Figure 7. A time series of temperature for a CTD located at 7 m depth just inside the Columbia River estuary at the Desdemona Sands light CORIE station for mid-may August In addition, temperature-salinity diagrams of the CORIE data for 2 relevant days (22 June and 20 August) corresponding to the field studies are also presented. The triangles represent a coastal station within 20 km of the river mouth at close to the same time. the same as observed at salinities of 5 10 within the estuary (Figure 9e). [23] The RISE-3W data from 18 to 20 August 2005 are from an upwelling-active phase and an intense spring tidal cycle. The winds had been strong and from the north almost continuously for the previous month with only a few 1 day breaks (Figure 3b). Figure 10a shows that warm river water of 21.5 C was mixing with cold (9 C), high-salinity ( 33) coastal seawater to form a plume with a salinity of and a temperature of C. Figure 10b shows that river water of 150 mm silicic acid was mixing with cold, nutrient rich seawater with 35 mm silicic acid to form a plume of salinity 20 with 80 mm silicic acid. The plot of salinity versus nitrate (Figure 10c) shows river water with 10 mm nitrate mixing with cold, nutrient rich water with 24 mm nitrate to form a plume with a salinity of 20 and a nitrate 11 of 23

12 Figure 8. A series of property salinity plots for data collected on the RISE-2W cruise on 5 7 June 2005 within the estuary (gray diamonds), within the core of the near-field plume (gray circles), and in a vertical profile of nearby coastal seawater (white triangles). This was during a downwelling-relaxation phase and warm, relatively low-salinity, nutrient-depleted coastal seawater was observed at depths of approximately 10 m and was the water entrained together with the river water to form the plume. 12 of 23

13 Figure 9. A series of property salinity plots for data collected on the RISE-2W cruise on June 2005 within the estuary (gray diamonds), within the core of the near-field plume (gray circles), and in a vertical profile of nearby coastal seawater (white triangles). This was during a downwelling-relaxation phase and warm, relatively low-salinity, nutrient-depleted coastal seawater was observed at depths of approximately 10 m and was the water entrained together with the river water to form the plume. concentration of mm, with 80% of the nitrate in the plume coming from the coastal seawater source. [24] The upwelling-active phase resulted in cold, highsalinity, nutrient-rich, low-oxygen coastal seawater with elevated dissolved Fe concentrations of nm being drawn into the Columbia River estuary to mix and form the plume. Figure 10d shows high dissolved iron concentrations in this near bottom water at depths of m being mixed 13 of 23

14 Figure 10. A series of property salinity plots for data collected on the RISE-3W cruise on August 2005 within the river (white circles), estuary (gray diamonds), within the core of the near-field plume (gray circles), and in a vertical profile of nearby coastal seawater (the dissolved Fe and Mn data are plotted for the subsurface coastal seawater end-member only (white triangles)). This was during an upwelling-active phase and cold, relatively high-salinity, nutrient-rich coastal seawater was observed at depths of approximately 10 m and was the water entrained together with the river water to form the plume. with estuarine water with dissolved iron on the order of 10 nm. The plume water at a salinity of 20 has a dissolved Fe concentration of nm. The dissolved Mn concentrations in the plume were elevated and ranged from 150 to 250 nm with a value of 230 nm at a salinity of 20. These high dissolved Mn concentrations in the plume fell along a mixing line between the coastal seawater with concentrations on the order of 100 nm and estuarine water at a 14 of 23

15 Figure 11. A time series of temperature for a CTD located at 7 m depth just inside the Columbia River estuary at the Desdemona Sands light CORIE station for mid-may August of In addition, two temperature-salinity diagrams for the CORIE data for 2 relevant days (26 May and 29 June) are also presented. salinity of 10 with concentrations of dissolved Mn of roughly 330 nm (Figure 10e). These concentrations are more than an order of magnitude greater than observed in the Columbia River and suggest a major source of dissolved Mn coming from the resuspension of estuarine sediments during the intense spring tide conditions Data From 2006 [25] Cruise data were obtained in May during RISE-4W. For the first half of May 2006, the system had been in an upwelling-active phase (Figures 3c and 11). MidMay was characterized as a downwelling-relaxation phase with generally strong downwelling winds from the south 15 of 23

16 Figure 12. A series of property salinity plots for data collected on the RISE-4W cruise on May 2006 within the river (white circles), estuary (gray diamonds), within the core of the near-field plume (gray circles), and in a vertical profile of nearby coastal seawater (the dissolved Fe and Mn data plotted are for the subsurface coastal seawater end-member only (white triangles)). This was during a downwelling-relaxation phase and warm, relatively low-salinity, nutrient-depleted coastal seawater was observed at depths of approximately 10 m and was the water entrained together with the river water to form the plume. 16 of 23

17 Table 1. Oceanographic, Tidal, and River Conditions for the Different Sampling Periods During the RISE Program Dates and relatively short periods of only a day or so of relatively weak winds from the north. During this downwellingrelaxation phase plots of salinity versus temperature at the Desdemona Sands CORIE station (Figure 11) indicate that the coastal seawater that was mixing with the Columbia River water was warm (12.5 C), low-salinity seawater. May 2006 was also a period of relatively high river flow (Figure 2a). [26] The RISE-4W data from 24 to 28 May 2006 are from a downwelling-relaxation phase, and during an intermediate tidal cycle. During the sampling on 25 May, river water of 15 C was mixing with relatively warm (12.5 C) coastal seawater to form a plume of salinity15 25 and a temperature of C (Figure 12a). Figure 12b shows that river water of 215 mm silicic acid was mixing with warm, nutrient depleted seawater with <3 mm silicic acid to form a plume of salinity 20 with 80 mm silicic acid. The plot of salinity versus nitrate (Figure 12c) shows river water with 12 mm nitrate mixing with warm, nutrient depleted water with <1 mm nitrate to form a plume with a salinity of 20 and a nitrate concentration of 4 mm. In this case, the nitrate in the plume had only a river source. [27] Figure 12d shows low dissolved iron concentrations of only a few nm in this warm nutrient depleted subsurface water at depths of 5 20 m being mixed with relatively high dissolved iron water within the estuary. It appears that during the high-flow conditions a much greater fraction of the river dissolved Fe was making it through the lowersalinity portion of the estuary. The plume water at a salinity of 20 had 13 nm dissolved iron. The dissolved Mn concentrations in the plume ranged from 50 to 150 nm with a value of 100 nm at a salinity of 20. The Mn concentrations observed in the plume were greater than observed in the warm, nutrient poor seawater and higher than observed in the river (Figure 12e). The highest dissolved Mn concentrations within the estuary during these high river flow conditions were observed at salinities of Discussion Upwelling/ Downwelling Tide Conditions River Flow (m 3 /s) Pre-RISE 2 3 Jul 2004 Upwelling Intense spring tides Low (4000) RISE-1W Jul 2004 Downwelling Neap tides Low (4000) RISE-2W 5 7 Jun 2005 Downwelling Neap tides Intermediate (7000) Jun 2005 Downwelling Low neap tides Intermediate (6000) RISE-3W Aug 2005 Upwelling Intense spring tides Low (4000) RISE-4W May 2006 Downwelling Intermediate tides High (12,000) [28] Table 1 presents a summary of the upwelling/ downwelling, tides, and river flow for the various field sampling efforts. Table 2 presents a summary of the chemical characteristics of the Columbia River and coastal seawater end-members during the various sampling intervals, and Table 3 presents a comparison of plume chemistry at a salinity of 20 during each of the RISE sampling periods. Although the salinity of the core for the near-field plume varied from 13 to 20, with lower salinities observed during neap tides and higher river flows and higher salinities observed during spring tides and lower river flows, a salinity of 20 was chosen for comparing plumes in Table Columbia River Fresh Water End-Member [29] The temperature, nitrate and silicic acid concentrations in the Columbia River during the different sampling time periods were generally what was expected on the basis of the 10 year time series presented in Figure 2. During the May August time period of interest, the temperature of the river was lowest in May and June (15 16 C) and increased to values in excess of 20 C in July and August. This is also apparent in Figures 4, 7, and 11 where the daily upper temperature in the estuary (coincident with the low-salinity value) approached the river temperature on each ebb flow of the tides. [30] The nitrate concentrations in the Columbia River were higher in May and June and lowest in the July and August sampling interval. The July 2004 nitrate concentration of 4 5 mm was consistent with those observed during the last decade in Figure 2a. The nitrate concentrations of mm measured in the river in June and August 2005 and in May 2006, although high relative to the longer-term record, were consistent with nitrate concentrations reported by the USGS in the summer of 2005 and 2006 (Figure 2a). Silicic acid concentrations in the river were elevated and ranged from 150 to 215 mm, with the highest concentrations in May and decreasing through the summer to lower values in July and August consistent with the long-term trend (Figure 2b). [31] Nitrate and silicic acid both appeared to behave conservatively within the Columbia River estuary as evidenced by their linear relationship with salinity (Figures 5b and 5c through 12b and 12c). This conservative behavior is consistent with the short residence time (1 to several days) of water within the Columbia River estuary [Jay and Smith, 1990; Nash et al., submitted manuscript, 2008] relative to the low biological activity within the estuary during the summer ( and PRIMARY_PRODUCTION.pdf). Phosphate (data not shown) also behaved conservatively within the estuary. During RISE-2W and -3W when the nitrate concentration varied between 9 and 17 mm, the phosphate concentration covaried with nitrate, with a nitrate:phosphate ratio averaging 17.3, a value close to a Redfield ratio of 16. This river water, slightly deficient in phosphate relative to nitrate and a Redfield ratio of 16, mixed with coastal seawater that had an excess of phosphate relative to nitrate and a Redfield ratio. Thus, with respect to the requirement of macronutrients by phytoplankton, the summer time plume was nitrate limited. [32] Dissolved Fe concentrations in the Columbia River itself, or in estuarine waters with salinity less than 1, ranged from 23 to 210 nm. The highest concentrations observed in 17 of 23

18 Table 2. River Water and Coastal Seawater End-Member Characteristics Dates Temperature ( C) Silicic Acid (mm) Nitrate (mm) Dissolved Fe (nm) Dissolved Mn (nm) River Water Coastal Seawater River Water Coastal Seawater River Water Coastal Seawater River Water Coastal Seawater Pre-RISE 2 3 Jul RISE-1W Jul <3 5 < RISE-2W 5 7 Jun < < Jun <3 13 < River Water Coastal Seawater RISE-3W Aug RISE-4W May <3 12 < these waters occurred during the August 2005 sampling period when concentrations on the order of 200 nm were measured in the Columbia River samples. These are relatively low concentrations of dissolved Fe compared to that reported for rivers in eastern North America where higher concentrations of dissolved Fe exist associated with humic substances [Boyle et al., 1977]. Nothing is known about the seasonal or interannual variation in dissolved Fe in the Columbia River. In each of the sampling periods, a nonconservative behavior of dissolved Fe in the low-salinity region of the estuary was observed, with dissolved iron rapidly decreasing and being removed from solution in the salinity range of 0 5. This is consistent with concentrations of riverine dissolved Fe being associated with humic substances that tend to flocculate and precipitate from the dissolved phase as the river water begins to mix with the coastal seawater [Sholkovitz, 1976; Boyle et al., 1977; Nowostawska et al., 2008]. Sholkovitz and Copeland [1981] and Nowostawska et al. [2008] demonstrated that the anionic sites of the humic acids are particularly effective at complexation with the increased Ca 2+ and Mg 2+ in seawater, neutralizing their charge and rapidly allowing flocculation and precipitation of the humic acids and their associated complexed Fe. Generally, in the salinity range of 6 10 within the estuary, relatively low dissolved Fe concentrations were observed. In both sampling periods during July 2004 a concentration of 4 nm was observed at a salinity of 6, and in June and August 2005 a concentration of 10 nm was observed at salinities of 9. In the high-flow conditions of May 2006, however, the dissolved Fe was 25 nm in the salinity range of [33] In this case of high river flow observed in May 2006, Fe did not appear to be removed as efficiently at the salinity range of 0 5, with the river concentration of nm decreasing to only 25 nm in the low-salinity region of the estuary. The water in the Columbia River estuary has a short residence time because of the large river flow relative to the small volume of the estuary. The even shorter residence time of water within the estuary during the high-flow conditions may result in higher dissolved Fe still being present at a salinity of 5 10 (Figure 12d), or perhaps under the high-flow conditions the humic acid concentrations were decreased and, as a result, less flocculation occurred. Although the reasons for this are not fully understood, we did observe a high fraction of the riverine dissolved Fe in the estuary and it was a substantial source to the plume during high river flow during May 2006; whereas during low river flow conditions there appeared to be major removal of dissolved Fe in the low-salinity regions of the estuary and the riverine Fe was only a minor source to the plume. [34] Dissolved Mn concentrations in the Columbia River or estuary at salinities of less than 1 ranged from 12 to 95 nm, with concentrations of nm observed in the river itself. Once again, reliable data on seasonal or interannual variations in dissolved Mn in the river are lacking. Removal of dissolved Mn was not observed in the estuary, instead an input was observed that coincided with tidal amplitude changes (Figure 13), with higher concentrations during spring tides (as high as 330 nm (Figure 10e and 13)), when enhanced tidal energy results in greater suspended sediment [Sherwood et al., 1990; Jay et al., 1990; Jay and Smith, 1990] than during neap tides (concentrations as low as 15 nm, Figure 8e). E. Y. Spahn et al. (Particle resuspension in the Columbia River plume near field, submitted to Journal of Geophysical Research, 2008) suggest that as the tidal range increases toward the maximum of a spring tide, the estuarine turbidity maximum is supplied with particles from peripheral areas and then exported during the strongest spring tide ebb flows. Aguilar-Islas and Table 3. Near-Field Plume Characteristics at a Chosen Representative Salinity of 20 Dates Temperature ( C) Silicic Nitrate Dissolved Dissolved Acid (mm) (mm) Fe (nm) Mn (nm) Pre-RISE 2 3 Jul RISE-1W Jul RISE-2W 5 7 Jun Jun RISE-3W Aug RISE-4W May of 23

19 Figure 13. Dissolved Mn (nm) at plume salinity of 20 versus tidal amplitude (m) for the various sampling periods in the RISE program. The tidal amplitude data is from within the estuary at Jetty A. Bruland [2006] have suggested that dissolved Mn can either be mobilized from suboxic pore waters along with the coexisting sediment during spring tide resuspension events, or alternatively, that the increase in dissolved Mn could be via photo dissolution of manganese oxide coatings of the resuspended sediments. [35] Sherwood et al. [1990] observed a range of mg L 1 in the concentration of suspended particles within the surface waters of the Columbia River Estuary, with the lower suspended loads found seaward during neap tides, and the higher loads found in the middle portion of the estuary during spring tides. The average Mn content in suspended sediment during the summer in the Columbia River is 1.5 mg g 1 [Covert, 2002]. These values give a range of suspended particulate Mn of nm within estuarine surface waters. A rate of 4.9% h 1 for the photo dissolution of particulate Mn in estuarine water was obtained by Sunda and Huntsman [1994] during laboratory studies using 54 Mn radio labeled manganese oxide particles. A lower dissolution rate in seawater (2% h 1 ) was reported by Matsunaga et al. [1995] during laboratory-based experiments. These studies used full sunlight conditions, and lower particle concentrations, therefore a lower Mn photo dissolution rate is likely under natural conditions of changing irradiance, and percent light transmission in surface waters. Allowing for 8 h of intense sunlight, and using 0.5% h 1 or 5% h 1 as possible rates of Mn photo dissolution, yields a daily contribution in dissolved Mn to estuarine surface waters from photo dissolution of suspended sediment of 5 55 nm and nm, respectively, for the range of calculated suspended particulate Mn ( nmol/l) in the surface waters of the estuary. Thus, this photo dissolution process could be the source of the elevated dissolved Mn during the spring tide periods (Figure 13) Coastal Seawater End-Member [36] Although the upwelling-active phase was the dominant mode during the summer months (Figures 4, 7, and 11), only two of our sampling periods coincided with upwelling-active phases (2 3 July 2404 and August 2005 (Table 1)). During these two sampling periods, the coastal seawater end-member had a temperature of C, a silicic acid concentration of 35 mm, nitrate concentrations of mm, dissolved Fe concentrations of nm, and dissolved Mn concentrations of nm (Table 2). The other four sampling periods coincided with downwelling-relaxation phases. In these cases the temperature of the coastal seawater end-member was C, the silicic acid was <3 mm, the nitrate concentration was <1 mm, dissolved Fe was 2 3 nm, and dissolved Mn was nm. Thus, there was a dramatic variation in the chemistry of the coastal seawater endmember that mixed with river water to form the plume. This variability led to marked differences in the macro- and micronutrient chemistry of the Columbia River plume (Table 3). During upwelling-active phases the plume had elevated nitrate concentrations ranging from 16 to 19 mm (with 90% from the coastal seawater source), while during downwelling-relaxation phases the plume had nitrate concentrations ranging from 2.5 to 6 mm (with only the river being a source). [37] On the basis of historical nutrient data over the last decade for the Columbia River (Figure 2), the low nitrate summer data from 2004 appears to be representative of a typical or average year. In the summer of 2004 the nitrate concentration in the Columbia River plume of salinity 20 during an upwelling-active phase was 16 mm, in marked contrast to a concentration of 2 3 mm in a downwellingrelaxation phase. A plume formed during an upwellingactive phase with 16 mm nitrate (90% from the coastal seawater source) will have a far greater impact on productivity than a plume formed during downwelling-relaxation phases with only 2.5 mm nitrate. During the downwellingrelaxation phase in June 2005 the observed nitrate concentration in the plume of 5 6 mm was higher than normal (2.5 mm) because of the higher-nitrate concentrations observed in the Columbia River during the summer of [38] The tidal phase (e.g., spring or neap tides) also has an impact on the nitrate concentrations observed in the plume at a salinity of 20 during an upwelling-active phase. During spring tides there is more seawater drawn into the estuary and the resultant near-field plume has a higher initial salinity than during neap tides. As a result, during spring tide/upwelling-active conditions more cold, high-nitrate water is mixed with the river water to form a high-nitrate plume of salinity 20. During neap tide/upwelling-active conditions, less nitrate-rich seawater is mixed with the river water and the lower-salinity near-field plume (salinity of 13) resulting in a lower nitrate concentration. Thus, summertime plumes with the highest-nitrate concentration are formed during upwelling-active, spring tide conditions. Nash et al. (submitted manuscript, 2008) have shown how the magnitude of the river flow also influences the salinity of the plume with lower salinities during high fresh water river flow. Thus, summertime plumes with the highest-nitrate concentrations will be during upwelling-active, spring tide, and low-river-flow conditions. This will be a plume with a relatively high salinity (>20) and the majority of the nitrate provided by the upwelling seawater source. [39] Although silicic acid concentrations also varied greatly in the coastal seawater end-member (<3 35 mm), the concentrations in the Columbia River water are extremely elevated, even during the summer months ( mm), therefore the river source always dominated the 19 of 23